Belt charging system

HAL Id: jpa-00244328

https://hal.archives-ouvertes.fr/jpa-00244328

Submitted on 1 Jan 1977

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Belt charging system

M. Letournel, J.-C. Oberlin

To cite this version:

BELT

CHARGING SYSTEM

M. LETOURNEL and J.-C. OBERLIN

Centre de Recherches

_Nucléaires,

Université Louis

_Pasteur,

67037

_Strasbourg

_Cedex,

France

Résumé. 2014 Les courroies des accélérateurs M.P. ont rencontré de sérieuses difficultés

principale-ment au _sujet de la durée de vie. Différents _paramètres du système de _charge par courroie sont _analysés et certains d’entre eux sont _soulignés comme concourant à la détérioration de la courroie ou à la

modu-lation de tension. Dans cet ordre _d’idées, certaines

_expériences

sont actuellement en cours et seront

poursuivies afin de

_pallier

à ces limitations de courroie et de déterminer _également les

_{possibilités.}

Abstract. 2014 In

M.P. accelerators belts have encountered some difficulties _mainly about the belt

life time. In the belt _charging system, parameters are investigated and some of them are

_pointed

out as _being

_partly

_responsible either for belt deterioration or

_for

belt _ripple.

_Experiments

are and will be carried out to surround the actual belt limitations.

In many

places

after

_carrying

a lot of nuclear

_physic

experiments,

belts seem to have encountered

_big

diffi-culties,

specially

in M.P. machines. Belt

_charging

system

seems _very

_simple.

Corona needles or now

screens,

deposit charges

at the base of the machine and _carry them _up to the terminal electrode where

they

are

_picked

up by a screen. In order to double the

_possibility

of

_charging,

_negative

_charges

can be

drawn from the terminal electrode down to the base. All this has worked well for _{many years.} Besides of

some well known

_advantages

as

_simplicity,

a lot of

current,

transported

belt show another

_{characteristic,}

the

_{voltage ripple,}

well

_{analyzed by}

_many

_people

_[1,

2].

This

_ripple

is the modulation in terminal

_voltage

due to the balance between the

_{incoming charge}

and the

_{outgoing charge}

_tempered

_through

the RC of the terminal electrode. A corona stabilizer have to _cope with the free

_{running ripple}

in order to

_bring

it around 1 or 2 kV

_[3].

Everyone

knows the main characteristics of the

belt,

but is also aware of some disasters with M.P.

belt

_(Fig.

_1).

Belt life time is

_commonly

more than 10 000 hours in _{many Van de Graaff}

_machines,

but

_specially

in the M.P. machine it has turned out

that belt life time is the

_major

critical

_point,

belt duration

_{being generally everywhere}

of the order of a

few thousand hours or even 1 or 2 000 hours. Black

and tan belts have shown the same characteristic

deterioration _process. A ruined belt shows

everywhere

similar failures:

_long

holes

_{along long stripes}

either

completely punctured through

the belt or

_partially

punctured

to the belt carcass and on both

sides,

individual

_{pinholes through}

the

_belt,

_among or

besides

_large

mat belt surface areas. A

special

_paper is

devoted to this kind of deterioration and structure

_[4].

The _purpose of this _{paper is} to _try to

_investigate

some

FIG. 1. - Belt breakdown.

fondamental _{processes in} belt

_-charging

_system, some

problems,

or some new _aspects.

1. Belt - Surface and bulk contact electrification

-Ripple.

- The belt is made of

a cotton carcass

bet-ween two

_layers

of rubber. This rubber has to _carry

charges

and in static

_electricity

there are three

elec-trification processes :

1 - Contact electrification

(including

tribo and field induced

_{electrification) ;}

2 - Corona

charging;

3 - Polarization in electric fields.

There is a

_strong

evidence that contact

electrifica-tion

_plays

a role in the contact between belt and other material as

_pulley

and

_guides.

In

fact,

it is

probably

an

_important

_component of the belt

_voltage

ripple.

The rubber

_coating

of the belt is a

poly-mer which is able to be

_{strongly charged through}

1384

contact to a metal. On

figure

2 we see the

_experiment

of Volta

_[5]

_showing

transfer of

_charge

after contact

between one

plate

of Zi and one of Cu. The contact

charging

metal insulator and

_specially

metal

_polymer

is

_highly

effective and has been studied

_recently

_by

many

people.

FIG. 2. - Volta’s

experiment.

On

_figure

3 Robins

_[6]

shows

_{charges acquired}

by polymers

after

_repeated

contact with a chromium

sphere.

We can see a kind of saturation with the

charge

densities

_according

to the number of contacts.

FIG. 3. -

Examples of _{charges acquired} after _repeated contact.

Contact

_charging

is

_{accomplished mainly by}

the

trans-fer of electrons and ions.

_Depending

on the nature of the material in _contact, the surface condition and.

external factors such as an electric field

_traversing

the

interface

a

_charge

transfer will occur. If the materials

become

_separated

some, none or all of the

charge

may remain on the

_respective

surfaces.

This

_charge

transfer has been

_{explained by}

Davies

_[7]

and Bauser

_[8]

and found to be

proportion-nal to the difference in the electron work functions of the

_contacting

materials

_{(Fig. 4).}

And the solid state

FIG. 4. -

Example of _{charge density against} work function.

Physics approach

has

_given

an

_explanation

similar

to that for

_inorganic

semiconductors. In static elec-trification of solid insulators

_by

metal _contacts, electrons are transferred from a metal to an insulator or reverse

_depending

on the distribution of the _energy

levels involved.

_Figure

5 shows the

_energy

band

_[9]

FIG. 5. - Contact between metal and polymer a) before contact

b) after contact.

schema for a metal and an insulator before and

_during

contact. The

_{charge [10]}

densities

_{produced by}

çontact

depend

on the nature and also on the _pressure

of the

contacting

materials. Cleves has shown the

strong

dependance [10]

of the

_{charge density}

with the

humidity.

And

it seems _very useful to think about a

fan or a

drying

device inside the tank to accelerate the

removal of

moisture,

just

after

_closing

the tank. In an

and the

_charge

transfer is _very

_{significant. Charge}

densities of 10-9

_C/cm2

or more can be

_generated

when a metal is

brought

into contact with an insulator

sample.

The electron transfer is instantaneous but it

depends

how the contact area is defined and in

prac-tice build _up of

_charge

is made after several contacts

depending

on how the contact is made and also

depending

on the nature and the cleanliness of the

surface,

and the contact _pressure

_[10].

_Mordhage

has shown the

_dependance

of the

_{charge density}

in contact

with an

_applied

electric field

_{[11]. Polymers}

absorb water which may

modify

the electronic structure of the

_polymer

and determine a _strong

_dependance

on

the

_charge

transfer. An

_experiment

that we have

car-ried out with Coste and

_Pechery

on a Van de Graaff belt

_sample

in 30

_%

relative

_humidity

has shown

negative charge

transfer of the order of 2

_nC/cm2

at

each contact, 10 contacts

_giving

15 x 10-9

_C/cm2.

These

_experiences

are _very well

reproducible

but

values are different from a

_sample

to another or even

from one

_place

to another on the same belt

sample .

Humidity plays

a

_major

role.

It is

_long

known that

_{rubbing insulating}

mate-rial

ône

on the other or on

metal,

make them to

get

charge

of a certain.

_sign.

Triboelectric lists or series

of different materials are in many books

[8],

and belt

do not avoid such a

_phenomena,

which is _very

compli-cated.

_Slipping

belts,

high

or low

_{ripple depending}

on belt

_tension,

_{might imply}

a _component

_coming

from this

_{negative charge}

transfer onto the belt. From the mechanical

_point

of view of a _pure contact and

_{separation avoiding}

_any

_sliding,

the

charge density

is

_proportional

to the difference in work function of

_the

_contacting

_partners. As soon as

gliding

or friction occurs, the

phenomenon

is far

more

_complex

and less

_{reproducible.}

And what is

contact and what is friction? Studies of the electrosta-tic

_{charges developed by}

friction under well controlled conditions

_specially

for relative

_humidity

have been done

_by

Coste and

_{Pechery (1)}

with an

_aparatus

where

there is a belt

_rotating

on rollers and

_gliding

is assured

by

one of the rollers which can tum as the same

_speed

or

_speed

_up or down or

_stopped.

A certain _pressure

can be

_applied.

The remarkable results of thousands of

_experiments

of different materials is that it is

possible

to divide the results in three _{different types}

(Fig.

6).

1 -

Charge density

is

_growing

_up to a constant

level. II -

Charge density

goes

through

a

_maximum,

then

_decreases,

_{changes polarity}

to obtain a new

equilibrium

state.

III -

Charge density goes through

a

maximum,

decreases to a constant level without

_{polarity change.}

(1) COSTE J. and PECHERY P. Private communication.

Experiences

are now

_being

carried out with a

belt and it seems that belt is of the _type I. It seems that

when we

_stop

the

_friçtion,

the normal

_polarity

and

charge

transfer occur in the samé _way as before but it

FIG. 6. - Charging friction effect _versus time.

is not

_always

the case and results are

_dependant

of the

surface itself and _many

factor.

For a

belt,

for

example

a M.P.

belt,

belt is driven

by

a drive motor whose drum is a

_designes,

with 17 grooves separated of 1 inch in order to let the _gas escape. This

pulley

is

_slightly

crowned from the two

last _grooves on each

_side,

down to the end. On the alternator

_{pulley crowning}

is also made on 2 inches on

each side. There the

_slippage

effect is obvious and is in the order of some

_cm/s.,

on the belt

_wings.

The

mechanical load of the belt is

_high

and

_obviously

not

the same

_along

the belt width because of the _grooves,

and also in

time,

according

to all the mechanical vibration modes of a

_running

belt. All these factors in

addition to the electric _up

_charge

field are

_part

of the

parasitic

charge

transfer,

and

_probably

of the

_ripple.

Some

_loading

on the

_normally

non

_grooved

alter-nator determines also the

_slippage

of the alternator

compared

to the belt. This

_phenomenon

leads to induce

a certain amount of

_{negative charges deposited}

onto

the terminal internal belt face. Due to the modulated load

_slippage,

it determines an erratic

_voltage

_ripple

sometimes

_higher

in SF6 than 30 kV and hard to

stabilize. The

_functionning

of the solution

_brought

through

the French connection

_(12]

_{(Fig. 7)}

was not

cléar

in the

_beginning,

but can be now

explained

by

induction of an

equal

amount of

_charge

of

_opposite

polarity

induced

_through

the second

_post

alternator

screen onto the extemal belt side to

_compensate

the internai

_parasitic

_charges.

2. Corona

_{charging process.}

- About the corona

charging

it is known that from a

_sharp

needle or a

very thin

wire,

the electric field around those

sharp

points may

increase

_beyond

the breakdown

_strength

’and corona

discharges

_appear. The _gas around the

1386

FIG. 7. - French connection device.

positive

or

_negative

_go

_along

the electric lines of force

and

_depending

on the field in the gap and the ratio of

point

radius to _{gap length,} the streamers start with very high speed and slow down to

_velocity

of some

100

_m/s

or more. Two

_things

must be

_pointed

out,

first for the

_layer

of

_charges

on the

_insulator,

the

extreme

_sensitivity

to the excess

_potential

on the

cathode,

second the

_strong

_dependance

of the current on the

_geometrical

characteristics of the wire or needle.

In other

words,

it would be not

_surprising

for a row of

needles or a screen to have the current

_coming

not

from each

_point

but

_mainly

from some of them.

3. Double

_{transportation possibility.}

- If

we

intro-duce an insulator into an electric field determined

_by

a constant

_voltage

U. After a certain

time,

due to

deposit

of ions of same

_sign,

the final state of this insulator is as shown on the

_figure

8a where the field

FIG. 8. -

a) Insulator between two plates. b) Natural double transportation.

between the insulator surfaces and the electrode is first

reduced,

then reduced to zero. In

_1953,

Gart-ner

_[13]

showed that the electrostatic machines could run under certain conditions

_along

the double natural

_charge

_{transportation. Figure}

8b shows this

principle.

For a certain

_field,

and

_mainly

if we

_get

ionization under the inductor

_plate,

there is

_deposit

of

negative charge density

-6,

charge

bound after a

certain time in the belt and never removed. The

char-ging

system works then that _{way only} in a case of

upcharge.

A

_{charge density}

2 a is _put on the belt

_by

corona

_discharge

for

_example

outside the belt and

removed

_by

a screen or needle in the terminal

elec-trode. The

_charge

-6 is built _up after some turns on

the internal belt side. This way of

charging

allows to carry a double

_{density. Only}

the

insulator,

belt

mate-rial encounters an extra strain. This double natural

charge transportation

is a

_possibility

at least

_partially

in the Van de Graaff machines.

_Experiences

of inter-nal belt

_charging

have

_proved

that

_charges

stay

without

_{being substantially}

removed

_by

the terminal

pulley,

and all the conditions are there to allow

partly

that kind of

_charging.

This process can lead to a non uniform

_pattern

of

_charges

bound on the

internal belt side due

_mainly

to the

_{parasitic charges}

and the non uniform drive motor

_geometry.

It can

, influence strongly

the non uniform

_deposit

of

_charges

by

corona _{process. To} avoid this

_possibility,

_try of belt

charge

removal can be done

_{simultaneously}

on both

sides after the terminal

_{pulley (Fig.}

_9).

FIG. 9. - Double discharge _screen.

4.

_Charge

removal. - Removal of

charges

from a

charge

insulator can be done

_through

different

pro-cesses but

_{mainly through}

the field

_{produced by}

the

charged

belt itself near the _top of

_{grounded points}

or

wires which ionized the _gas.

_Negative

ions are then

projected

on to the

_charged

belt in such a _way us to

neutralize the excess field at each instant. In the case

of the

_(Fig.

_10),

these ions are

_deposited

on a well

FIG. 10. -

Discharge mecanism of a foil due to an _approaching

earthed point: a) before the projection of ions,

b) during the projection of ions,

defined

surface,

the dimensions of which

_depend

on

the

_charge

density

and on the

_proximity

of _any conductors

_[14].

Whatever the

_speed

is,

a

_grounded

screen or serie of needles will

usually

discharge

the belt to an _average

charge density

of about 100

_pC/cm2

that means for the belt

_charge

current, in the order of 1

_J.lA

or less.

If there is

_enough

_space free of

conductors,

this eliminator works

_properly

but there are other means

to reduce this value if _necessary. As the initial

_speed

of these ions is _greater than

105

_cm/s

the

_discharge

follows the field variations almost

_{instantaneously.}

And in case of a H.V.E.C. belt _arrangement and in

order to _{prevent influence} from one side on the

other,

two screens with

_longwires

one on each

_side,

facing

each other on one cm distance can be added

after the terminal

_pulley,

with

_good

result to take care

of both sides

_(Fig.

_9).

To

measure the

_efficiency

of a

charge

eliminator Lôvstrand

_[15]

made a machine

which indicated the better residual

_{charge density}

results with

_{25 03BC}

wires and the

dependance

with the band

_velocity,

the

_{charge density}

and the nature of the eliminators.

Under the usual Van de Graaff machine

conditions,

as the

_charged

belt

_voltage

can not be

_higher

than this determined

_by

the belt

_capacity

_itself,

with

_charge

density higher

than 4

_nC/cm2,

two

_neutralizing

_steps can be considered in order to

_prevent

a _spontaneous

discharge,

before the _proper removal. The first _step

must be in a metallic environment which

_keeps

the belt

_charge

_{capacity high}

_enough.

It means that the

first screen must be at a distance not _greater than

some 2 to 4 cm from the

_ground,

or in another hand

belt must be

_{protected by}

a

_{grounded plate}

not too far from the first screen.

5.

_Physical

_processes involved with belts. - A belt

,

has

to be treated as an insulator of a

resistivity

of the

order of 1012 Qcm. Then with no free electrons on

the rubber

_surface,

the cotton carcass of

_resistivity

of the order of

109

Qcm

_being

there from the electric

point

of view to allow a

_good

distribution of the

potential. Consequently charging

process, removal process and

charge

transportation

must be

_explained

with

this

_particular

view of an insulator.

First,

speak-ing

about the so called

_rearranging

of

_charges

on the

belt to

_{explain ripple}

or loss of

_charges

in normal

running

conditions,

it is

_impossible.

As

_{Zichy point}

out

_{[16], charges}

are

firmly

bound inside the belt

polymer

structure and it is

_impossible

for them to

move _except

_through

_{gas ionization determined in} a

fixed field

_{configuration}

by

the belt

_{holding capability.}

Second,

the

_exchange

of belt

_charges

or the

_neutralizing

of these

_{charges proceeds,}

under normal conditions

only through

the

incoming

of ions of

opposite

polarity

created in the _gas. Electric lines of force

_(Fig.

₁₁₎

coming

from the _top surface of a

_charged

belt are

strongly

influenced

_by

the

_ground

and directed towards the

_nearby

conductors. The field

configura-FIG. 11. - Electric lines of forces configuration.

A. Charged belt on _pulley with a screen.

B. _Çharged belt alone with a screen.

C. _{Charged belt,} screen and nearby conductor.

D. _Charged belt on _plate, screen.

E. Internal _charged belt.

tion in the

_pressurized

_gas

_depends

on the

geometrical

position

of the

_nearby

conductors and other neutral

or

_{charged objects (C).}

_Then,

to act

_{properly, charge}

removal determined

_by

the belt

_charge

field itself must be

_accomplished

far

_enough

from conductors

_(B),

otherwise the field created on the screen is to low to

allow a

_complete

neutralization

_(D).

On the _contrary

this

_phenomenon

also

_{explains why}

it is

_impossible

to

remove _any

_charge

at all even with a

_grounded

screen

from the external side of a

_charged

belt

_right

onto a

metallic

_{pulley (A).}

_By

lack

_{of field,}

we must create an

additional electric field in order to neutralize the

charge

or

eventually

to

_charge

with the other

_polarity,

by

incoming

ions of

_{opposite polarity}

which is the

case of down

charging.

The field must be _strong

enough

only

where

_charging

or

_charge

removal has to occur. This

explains

why

under normal

conditions,

charges

can _stay

_during

an infinite time without

being

removed and are the basis of the terminal

voltage

repetitive

ripple

and of the

_possible

double natural

transportation. Particularly charges

bound in the belt inside surface are not removed

_by

contact on metallic

pulleys (E).

But,

in another

_hand,

we can think also

about the use of this

particularity

to

_imprint

durable

1388

6.

_{Charge transportation.}

- Another fundamental

problem

is the

_{charge transportation through}

the total

structure. It is

_{governed essentially}

_by

the _concept of

charge

belt

_capacity,

the deduced

_potential,

and _gas

rigidity.

The main _aspect of the

_{charge transportation}

is studied elsewhere

_[4],

and

_gives

an

_explanation

for the M.P. belt deterioration. About the other

_possible

belt failures

_(2),

_{voltages tracks,}

surface or volume break-downs are

_mainly

due to a

_{malfunctioning}

or

impro-per oimpro-peration somewhere. Over

_voltage

or over

char-ging

of the

belt,

poor charge spray or

_collection,

inducting inhomogeneity,

wet _gas or lack of

condi-tionning

are the

_probable

causes of failures. The

importance

of belt moisture removal before normal

operation

has to be

_emphasized.

If

_charges

move on a

belt

_surface,

it is because of too

_high

a

_tangential

field and a reduced surface

resistivity

due for

example

to

moisture.

7. Modifications in structure. - In

May

1976,

cha-sing

the kind of trouble due to M.P. belt deterioration which was

_patent

at this time

through

a current

leakage

on the internal structure section 8 and

follo-wing

the Van de Graaff and

_Trump

_concept

of

controlling voltage

on the

_belt,

we took all _spacers

and control rods _{away. From} the drive motor to the terminal

_pulley,

the belt was not allowed to touch

anything

except

the four screens. So the outside

guides

and bars were

_positioned

to follow as best we could

the

_{practical running}

curve of the belt which is

some-thing

as a _catenary curve.

_Spacers

and

_gradient

rods

were

_roughly

at a distance of 1 inch from the _up run

and

_1/2

inch from the down run which was not

_charged.

The mechanical

_ripple

was in the order of 1 cm for the

belt

_wings

but not more than 5 mm for the middle. To control from

outside,

we added 3

_{lamps (Fig.}

₁₂₎

in

FIG. 12. - Half

open structure without rollers.

. the terminal in vertical

setting

which we watched

through

a mirror

_system.

We ran with this structure

with

_exactly

the same

_voltage

characteristics _up to

13 MV and with no difference at all from

_previous

behaviour.

This half _open structure is fine. Since then in order

to stabilize and smooth the

machine,

we have installed insulated rollers

_{(Fig. 13)}

in dead

sections,

termi-(2) H.V.E.C. Private communication.

FIG. 13. - Half

open structure with some rollers.

nal electrode and

_high

_energy

_position.

Some of them

are coated with rubber and

_they

all run at

_floating

potential (Fig.

14).

We still had the same

voltage

FIG. 14. - Rubber coated roller.

characteristics and the rollers run

_perfectly

well.

_They

will be used in the future to better define and stabilize belt

_charges.

Figure

15 and

_figure

16 show the alternator and drive motor covered with two kinds of

_rubber,

and with holes which evacuate the SF6 _gas towards the ends.

_They

have been tried

_successfully

in the machine

as

regards voltage.

Besides the

good

friction

coeffi-cient,

the reason to have rubber is that the best non

electrification transfer contact takes

_place

between two

material of the same nature. Antistatic rubber can also

be used in order to eliminate

_{parasitic charges.}

But we can not

_keep

them like

that,

because of a belt

trans-versal abrasive

_{displacement,}

on the antistatic black

rubber.

_{They correspond}

to the final

_design

and we

have decided to

_replace

this rubber

_by

an identical

sleeve of

metal,

just

thicker.

As far as the

_stability

is

_concerned,

the same

philo-sophy

as for thé

_pelletron

must be

_considered,

to

control the total terminal

_charge

bilan. We must

_try

to

cope with the total belt

charge including

the

parasitic

charges.

A

_good

direction is

_certainly

to think as

Langsdorf [17]

of

_treating

belt

_charging

and

dis-charging by

induction. He mounted a belt

_charging

belt face and

_brought

the terminal free

_running

ripple

in the order of 300 or 500

_WBlythe

_[18]

presen-ted a similar induction device with

corona

wires. He

succeeded to control

_charge

_density

for

_charging

and

discharging

with a _very accurated _way. His device

does not

_operated

in normal Van de Graaff

running

conditions,

but

_joined

to our

rollers,

with some

changes

it can

bring

a

_great

improvement

to control the

_charge

_dénsity

and therefore the terminal

voltage

ripple.

. References

[1] WEITKAMP W. G. and SCHMIDT F. H., Nucl. Instrum.

Methods 122 (1974) 65.

[2] VERMEER A. and STRASTERS B. _A., Nucl. Instrum. Methods. 131 (1975) 213.

[3] THIEBERGER P., LINDGREN R., MCKEOWN M. and

WEGNER H. E., Nucl. Instrum. Methods 122 (1974) 527. [4] LETOURNEL M. and BELT M. P., Revue Phys. Appl. 12 _(1977). [5] RUPPEL W., (Dechema-Monographien) Nr. 1370-1409, 1973,

321.

[6] ROBINS, ROSE A. C., INNES and LOWELL J., Inst. Phys. Conf.

Ser _{27 (1975)} 115.

[7]DAVIES D. K., Inst. Phys. Conf. Ser. 4 (1967).

[8] BAUSER H. _{(Dechema-Monographien)} Nr. 1370-1409, 1973,

11.

[9] KRUPP H. Inst. Phys. Conf. Ser. 4 (1967) 1.

[10] CHALLENDE R. F. Advances in Static Electricity Vienne Auxilia-Brussels 1970.

[10*] CLEVES A. C., Inst. _{Phys. Conf.} Ser. 11 ₍₁₉₇₁₎ 226.

[11] NORDHAGE F. and BÄCKSTRÖM G. Inst. _{Phys. Conf.} Ser. 27 (1975) 84.

[12] Tandem accelerator Conference 1974 Rehovot - Israël. [13] GARTNER E., R.G.E. Février et mars 1953 t. 62 71-86,136-154. [14] BERTEIN H. Inst. _{Phys. Conf.} Ser. 4 ₍₁₉₆₇₎ 11.

[15] LOVSTRAND K. G. Inst. Phys. Conf. Ser. 27 (1975) 246.

[16] ZICHY E. L. (Dechema-Monographien) Nr. _{1370-1409, 1973,} 147.

[17] LANGSDORF A. Proc. Int. _Conf. on the Techn. Elect. Acc.

Daresbury (1973) 255.